I made some proof of concepts with ServiceMixm including a weekend running.
environment:
- 1 test machine (1Gb RAM)
- 1 oracle 8i db in other machine
ServiceMix 3.2 snapshot
Java 1.6.0
Prototype:
- 1 J2SE application that generates 50 records per second into the database
- 1 ServiceMix instance running my prototype code
Design:
- the prototype uses 3 lightweight components:
[input binding] - this component read each new record from DB and generates
a normalized message.
public class GansJdbcDataImporter extends PollingComponentSupport{...}
[service engine] - this component read the message generated by the reader
and create a new normalized message with some content (simulating a business
process)
public class ThresholdEngine extends OutBinding {...}
[output binding] - this component reads the message from the engine and then
writes some log into a file.
public class SmsHttpConnector extends OutBinding{...}
All components writes a timestamp in a monitor table - in order to give me a
glimpsy about the performance of the router and also to give me a chance to
followup the messages.
Test results:
the J2SE application generated 2.5 million records into the database
the JBI container processed 85 thousand messages - and then had problems of
memory.
Evaluation:
it seems the components are not removing the messages from the router after
processing it. Other curious problem is that the stack trace show excess of
threads. I supose there is a pool taking care of the number of threads and
components, so I cannot figure out why the system shutted-down after some
hours.
Possible error causes:
- wrong components code. The type of the components I didn't figured out
very well..
the stack trace:
> ERROR - PollingComponentSupport - Failed to schedule work:
> java.lang.OutO
> fMemoryError: unable to create new native thread
> java.lang.OutOfMemoryError: unable to create new native thread
> at java.lang.Thread.start0(Native Method)
> at java.lang.Thread.start(Thread.java:597)
> at
> edu.emory.mathcs.backport.java.util.concurrent.ThreadPoolExecutor.add
> IfUnderMaximumPoolSize(ThreadPoolExecutor.java:477)
> at
> edu.emory.mathcs.backport.java.util.concurrent.ThreadPoolExecutor.exe
> cute(ThreadPoolExecutor.java:876)
> at
> org.apache.servicemix.executors.impl.ExecutorImpl.execute(ExecutorImp
> l.java:42)
> at
> org.apache.servicemix.components.util.PollingComponentSupport$PollSch
> edulerTask.run(PollingComponentSupport.java:163)
> at
> org.apache.servicemix.components.varscheduler.Scheduler$SchedulerTime
> rTask.run(Scheduler.java:108)
> at java.util.TimerThread.mainLoop(Timer.java:512)
> at java.util.TimerThread.run(Timer.java:462)
> #
> # An unexpected error has been detected by Java Runtime Environment:
> #
> # java.lang.OutOfMemoryError: requested 1024000 bytes for GrET in
> C:\BUILD_AREA\
> jdk6\hotspot\src\share\vm\utilities\growableArray.cpp. Out of swap space?
> #
> # Internal Error (414C4C4F434154494F4E0E494E4C494E450E4850500017),
> pid=12324, t
> id=12400
> #
> # Java VM: Java HotSpot(TM) Server VM (1.6.0-b105 mixed mode)
> # An error report file with more information is saved as
> hs_err_pid12324.log
> #
> # If you would like to submit a bug report, please visit:
> # http://java.sun.com/webapps/bugreport/crash.jsp
> #
>
--
View this message in context:
http://www.nabble.com/out-of-memory---JVM-shutdown-tf3387721s12049.html#a9429944
Sent from the ServiceMix - User mailing list archive at Nabble.com.
